学位論文

ウェーブレット変換を用いた低線量X線画像に対する量子雑音除去とコントラスト強調

jpn

2010-03-24

Not Applicable (or Unknown)

2010-03-24

修士(工学)

工学研究科情報電子工学専攻; 指導教授: 尾川浩一

An important issue in diagnostic imaging is to acquire high quality images with a small dose. However x-ray images acquired with a small dose are affected by the quantum noise. In this paper, we propose a new method to yield high quality images acquired with a small dose by reducing the noise and enhancing the contrast of images adequately. The proposed method consisted of the following steps : (1) pre-processing of an original image, (2) making 16 images with shifting the position of all pixels by several pixels in the image, (3)decomposition of these 16 images with a wavelet transform and reducing Poisson noise with a modified BayesShrink method in the wavelet domain, (4) reconstructing 16 denoised images with an inverse wavelet transform and restoring the pixel position, (5) averaging of 16 images pixel by pixel in the spatial domain, (6) decomposition of the averaged image with the wavelet transform and enhancing the contrast in the wavelet domain, (7) reconstructing a denoised and contrast enhanced image with the inverse wavelet transform, and (8) post-processing of the obtained image. To confirm the validity of our proposed method we acquired x-ray images of a standard human body phantom with small doses of 1/12 and 1/25 of the standard dose，and applied our method to these low-dose x-ray images. The validity of the proposed method was confirmed in comparison with the results processed using conventional methods.

Thesis